AI for your business
AI consultancy for SMEs
Before you invest in anything, it helps to know what it will deliver. We go through your process and tell you where AI would make a difference, what that costs and where you are better off leaving it alone.
At Darvis, AI consultancy is an assessment, not an implementation project. We join the people who do the work, map out where reading, sorting and retyping takes place, and deliver advice with the time saved, the build effort and the risk for each candidate. You also get an answer on what is allowed with your data. The advice stands on its own: you are under no obligation to have it built by us.
Does this sound familiar?
Most businesses are not short of possibilities, but of a reason to choose between them.
“Everyone says we need to do something with AI”
The push comes from the board, a customer or a trade fair. Nobody has worked out what it should actually deliver for you.
“We have tools, but no results”
Subscriptions have been taken out and colleagues are trying all sorts of things. Yet nothing has changed in how the work gets done.
“We do not know what is allowed”
Can customer data go into an AI service? What do we need to record? Without an answer, everyone just does their own thing.
“Every proposal promises the same”
Companies offering AI tell you what is possible. What it would save you and what it will cost remains vague.
What you get
Advice that leads to a decision
Not a presentation about what is possible, but a list of what there is to gain for you, in order of what it delivers.
Walking through your process
We join your team and map out where reading, sorting, retyping and writing takes place. That is where there is something to gain, and it is rarely where people expect.
A prioritised list
Each candidate gets an estimate of the time saved, the build effort and the risk. That lets you choose, even if you choose to do nothing.
What is and is not allowed
Which data may leave the building, which agreements you need with a supplier and what you need to record. Concrete, and applied to your situation.
Advice you can act on
You receive a report with the findings and a proposal for the first step, including what it would cost. Even if you have that step built by someone else.
Already know where you want to start? Then you can have that step built straight away; how we do that is explained under AI automation for businesses.
Scope
What we do not do
We do not train our own language models and we do not sell an AI platform. What we do is determine which work in your business lends itself to automation, which of it can be done with ordinary software and which needs a model, and then build that into the systems you already use. That is a deliberate choice: for the vast majority of questions, the gain lies in combining existing services with your own data, not in building something new at model level.
We also do not give advice about anything we could not build ourselves. If the assessment produces an outcome that falls outside our field, the report simply says so.
How it works
From introduction to decision
It starts with a one-hour conversation. Often that is already enough to decide whether an assessment makes sense.
Introduction
A conversation about what you do, how the work runs today and where the question comes from. No obligation, and usually enough to decide whether an assessment makes sense.
Joining your team
We follow the process with the people who carry it out. Not as it is described, but as it actually happens.
Working it out and costing it
For each candidate we estimate the time saved and the effort, and check whether it is responsible with your data.
Discussing and deciding
You receive the report and we go through it together. Then you decide what happens next, and whether we build it or someone else does.
Frequently asked questions about AI consultancy
What does an AI assessment actually deliver?
A report showing the places in your process where AI would make a difference, with an estimate of time saved, effort and risk for each, and a proposal for the first step including its cost. That lets you make a well-founded decision, even if the outcome is to wait a little longer.
Do you ever advise doing nothing?
Yes, and it happens regularly. Often the gain turns out not to lie in AI but in an ordinary integration or in cleaning up double data entry. When that is the case we say so, even if it means a smaller assignment for us.
Do we have to have it built by you afterwards?
No. The advice is yours and stands on its own, with enough detail for another party to carry it out. We are happy to build it, but the assessment is not there to force that.
How much of our time does the assessment take?
Allow for a one-hour conversation and half a day alongside the people who do the work. We do the rest ourselves. The longer we are allowed to observe, the more concrete the advice becomes.
Is it acceptable to send our customer data to an AI service?
That depends on which data and which service, and it is exactly one of the questions the assessment answers. There are business terms under which your input is not used for training, and there are steps for which you do not need to send any personal data at all.
We are a small business: is this for us?
Very much so, because the candidates are then easy to oversee. In a small business it is usually about one or two tasks that come back every day, and that is exactly the kind of step that pays for itself quickly.
If the gain turns out to lie not in AI but in ordinary automation, read how we approach that under business process automation.
What would AI save you?
One conversation of an hour and you will know whether there is something to gain. You will receive a response to your enquiry within 24 hours.